Multilevel Analysis in Public Health Research

Abstract

Key Words methods, contextual effects, random effects, social epidemiology, ecologic s Abstract Over the past few years there has been growing interest in considering factors defined at multiple levels in public health research. Multilevel analysis has emerged as one analytical strategy that may partly address this need, by allowing the simultaneous examination of group-level and individual-level factors. This paper reviews the rationale for using multilevel analysis in public health research, summarizes the statistical methodology, and highlights some of the research questions that have been addressed using these methods. The advantages and disadvantages of multilevel analysis compared with standard methods are reviewed. The use of multilevel analysis raises theoretical and methodological issues related to the theoretical model being tested, the conceptual distinction between group-and individual-level variables, the ability to differentiate "independent" effects, the reciprocal relationships between factors at different levels, and the increased complexity that these models imply. The potentialities and limitations of multilevel analysis, within the broader context of understanding the role of factors defined at multiple levels in shaping health outcomes, are discussed.

Unravelling the extent of inequalities in health across urban and rural areas: Evidence from a national sample in England

“…Multilevel models are particularly useful because they account for the clustering of similar individuals within areas (non-independence of observations within groups), the variation in health outcomes is attributed to differences between individuals and to differences between areas, and they allow for the assessment of the relative contribution of both individual-level and area-level characteristics on these two sources of variation (Diez-Roux, 2000).…”

Residential Segregation and Latino, Black and White Mortality in New York City

“…Specifically, because the mortality data were clustered within zip codes, it was possible that there was some intraclass correlation of deaths occurring in the same zip codes. [46][47][48][49][50][51] However, we noticed that most of the zip codes included in the analyses have few deaths, suggesting that the outcomes may be independent and thus, not necessitating multilevel analyses.…”

Mapping and Measuring Social Disparities in Premature Mortality: The Impact of Census Tract Poverty within and across Boston Neighborhoods, 1999–2001

“…14 With respect to statistical considerations, multilevel modeling using generalized linear mixed models has become an increasingly popular method for dealing with small area rate instability by Bsmoothing^over unstable rates. [16][17][18][19][20] This approach decomposes overall variation at multiple levels of spatial aggregation and models the contribution of Bplace^using random effects that are assumed to be drawn from a common underlying distribution. This effectively permits one to Bborrow strengthâ cross neighboring areas in the multilevel hierarchy to yield more stable place level estimates.…”